On the tenability of sceptical arguments